Facebook and #GE2015
Facebook I voted button
Like anyone else from UK on Facebook this morning, I saw the 2015 General Election I voted button that asks you to confirm if you have voted. I was interested to see the data it presents – a time series graph of number of voters who claim they have voted, and an age/gender distribution of the same – you will notice that my captions differ slightly from the app’s captions.
Time series of users of Facebook GE2015 app
Distribution of Voters app users by Age/Gender
When I checked out where the App came from, I was interested to see it was linked to the About My Vote site from the Electoral Commission, and I posted an enquiry about their and Facebook’s use of the data generated by the app. I will update this post with their response if and when I get it.
I got into a conversation with my son Dan Bell who is a Digital Marketing Consultant (and takes a different perspective on Facebook from me :). He showed me the dashboard of his Facebook Audience insights and so we checked out if you needed a Google Facebook Ads account to see that data – you don’t!
Facebook data on UK users with interest ‘UKIP’
The image is from my quick exploration of going to https://www.facebook.com/ads/audience_insights, making following choices:
Audience: Everyone on Facebook, changing country by selecting United Kingdom, deselecting USA, and typing UK Independence Party in as an interest.
What isn’t clear to me is how Facebook decided who was ‘interested’ in UKIP. I will be very interested if Nigel Farage fails to take Thanet South seat. Dan pointed out to me that it isn’t just who has liked their page as the UKIP page has only 464,000 likes yet the graph above relates to ‘700,000-800,000’ users.
It does seem poor that Facebook Audience Insights are coy about their methodology/ algorithm, given their concerns about privacy.
We built Audience Insights with privacy in mind. It surfaces aggregated information people already express on Facebook, along with information from trusted third-party partners — like Acxiom — through our partner categories targeting.
I can recommend having a play with this – and then you might like to go here and adjust your ad settings 🙂 https://www.facebook.com/help/568137493302217?hc_location=ufi
So how do you feel about these services? Have you any concerns? I do.
Chris Thomson (@cbthomson) May 7, 2015 - 3:29 pm
“What isn’t clear to me is how Facebook decided who was ‘interested’ in UKIP. ”
This fascinates me. The reasons why people might show what Facebook calls “interest” in something will be highly nuanced and yet all Facebook (and other behaviour data-based corporations) can do with it is crudely call it “interest”. How do they determine that I view a page because I sympathise with it politically or because it’s got a picture of someone I remember from school on it?
I guess when you’re working at such massive scales it doesn’t matter if you’re serving up ads that miss their target some of the time but it does make you question just how much venture capital is invested on the bet that the data that is collected is sufficiently understood to be effective.
But then, maybe all that matters to Facebook’s customers is where our attention is focused.
Personally, I’m quite pleased that there are parts of the human experience that algorithms can’t reach.
You’ve given me an idea for a blog post!
francesbell May 7, 2015 - 5:30 pm
These are two links from my recent stream
http://www.theatlantic.com/politics/archive/2015/04/when-cops-check-facebook/390882/
https://medium.com/ursium-blog/the-end-of-big-data-a-reasonable-internet-of-things-ff78a4428ed7 but I can’t find the one I was looking for that did some sort of analysis on open social media data and predicted personal characteristics for employers of their prospective employees.
What worries me is that we might be groomed by one set of services into accepting intrusion and surveillance of our public words and deeds by automated collection and analysis into accepting another service that feeds into human decision-making.
francesbell May 7, 2015 - 8:18 pm
Just adding a brilliant link via Paul Olivier deHaye http://thesocietypages.org/cyborgology/2015/05/07/facebook-fair-and-balanced/
Chris Thomson (@cbthomson) May 8, 2015 - 12:40 pm
Excellent links there, Frances. Thanks. I also echo you’re idea of “grooming” here. I’m not sure we as educators do a good enough job of helping learners to understand this and be critical of the tools they use. We tend to talk in terms of safeguarding learners by helping them protect themselves from other service users but we don’t address the relationship they have with the service provider and it’s implications. If full participation in a course relies on having a Facebook account, for example, do we help them to question using a service who’s business model doesn’t necessarily have their best interests at heart?
francesbell May 8, 2015 - 2:54 pm
Thanks Chris, this is great – going beyond my post:) We often use the term tool (I do myself) but I think we need to go beyond tools, and systems, to think about the lash ups, the messes the networks that writhe around what we might think of as ‘Facebook’. It can’t really be explained as a tool or even a service – it has arrived where it is (differently for you and for me) on a messy historic trajectory involving Zuckerberg , Venture Capitalists, advertisers, Facebook members, clickbait providers like viralnove, etc., etc.
I do think that questioning the assemblage that is Facebook is important in education as is challenging a belief that we can’t make a difference – not to make it a particular way but to shift it and even slightly reshape it.